Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Do you want to argue we should continue allowing it?
No; I'm objecting that there hasn't been adequate discussion about
this change of behavior.
BTW, if the rationale for the change is "ANSI compliance" then the patch
is still wrong. SQL92 says:
3) No <column name> of T shall be identified more than once. If the
<insert column list> is omitted, then an <insert column list>
that identifies all columns of T in the ascending sequence of
their ordinal positions within T is implicit.
5) Let QT be the table specified by the <query expression>. The
degree of QT shall be equal to the number of <column name>s in
the <insert column list>.
The patch enforces equality only for the case of an explicit <insert
column list> --- which is the behavior I suggested in the original
comment, but the spec clearly requires an exact match for an implicit
list too. How tight do we want to get?
In any case this discussion should be taking place someplace more public
than -patches.
regards, tom lane